CVE-2026-84286

OOB Read/Write via Negative Array Index in exllamav3_ext Kernel Table

Published Oct 12, 2026 CVSS 3.1 · 9.8 Critical EPSS < 1% KEV · Not listed
Description
ExLlamaV3’s EXL3 loader uses an untrusted, model-supplied index to select a CUDA kernel without first verifying that the index is within bounds. A malicious model can force the code to read and invoke an out‑of‑range function pointer, resulting in a reproducible crash. Under specific conditions involving certain host memory layouts, it may be possible for an attacker to exploit this vulnerability to achieve controlled indirect kernel dispatch during model load.

Analysis

AI-generated

Technical impact

ExLlamaV3’s EXL3 loader uses an index supplied by the model itself to choose a CUDA kernel without verifying that the index is within bounds. If a malicious model supplies a negative or otherwise out‑of‑range index, the loader can read and call a function pointer that does not belong to the allowed kernel table, which can cause a crash and, under particular host memory layouts, may allow the attacker to trigger a controlled indirect kernel dispatch during model load. The resulting vulnerability can compromise the confidentiality and integrity of the system by executing code in kernel space, and can lead to a denial of service through a reproducible crash.

Affected Systems

Turboderp-org’s ExLlamaV3 is affected. No specific version information is provided, so any install that includes the EXL3 loader is potentially vulnerable.

Risk and Exploitability

The lack of a bounds check makes the issue technically high severity, but exploitability depends on the attacker’s ability to supply a malicious model and on the host memory layout. With a CVSS score of 9.8, the risk is weighed as critical. EPSS is <1%, and the vulnerability is not listed in CISA KEV, so current public exploitation activity is unknown. The likely attack vector involves an attacker delivering a crafted model to the application, which then loads the model locally or via an API. If the application accepts untrusted models, this presents a local or remote kernel‑mode code‑execution risk.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade ExLlamaV3 to a version where the kernel index is validated against the bounds of the kernel table
  • Validate any model‑supplied indices to ensure they are non‑negative and within the allowed range before they are used to select a CUDA kernel
  • If a patch is not yet available, deny or sandbox the loading of untrusted models until bounds checking can be enforced

Generated by OpenCVE AI on October 19, 2026 at 13:43 UTC.

Tracking

Sign in to view the affected projects.

Advisories

No advisories available
History
Type Values Removed Values Added
Weaknesses CWE-125
CWE-20
CWE-787
Metrics cvssV3_1

{'score': 9.8, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H'}

ssvc

{'options': {'Automatable': 'yes', 'Exploitation': 'none', 'Technical Impact': 'total'}, 'version': '2.0.3'}

Type Values Removed Values Added
First Time appeared Turboderp-org
Turboderp-org exllamav3
Vendors & Products Turboderp-org
Turboderp-org exllamav3
Type Values Removed Values Added
Weaknesses CWE-126
CWE-188
Type Values Removed Values Added
Description ExLlamaV3’s EXL3 loader uses an untrusted, model-supplied index to select a CUDA kernel without first verifying that the index is within bounds. A malicious model can force the code to read and invoke an out‑of‑range function pointer, resulting in a reproducible crash. Under specific conditions involving certain host memory layouts, it may be possible for an attacker to exploit this vulnerability to achieve controlled indirect kernel dispatch during model load.
Title OOB Read/Write via Negative Array Index in exllamav3_ext Kernel Table
References
Weaknesses

Sources

Status
PUBLISHED
Assigner
certcc
Published
Updated
Oct 19, 2026, 11:04 AM
Reserved
Sep 1, 2026, 3:47 PM
Status
Awaiting Analysis
Published
Oct 12, 2026, 5:17 PM
Modified
Oct 19, 2026, 12:17 PM